Document Xo7ZXjxV463mnZp1Gdnz24ax
496 Biochemistry 1990. 29, 496-504
Styring, S,, & Rutherford. A. W. (1987) Biochemistry 26, 2401-2405.
Tamura. N., & Martin. G. (1985) Biochim. Biophys. Acta 809. 245-257.
Teo. B.-K., & Lee, P. A. (1979) J. Am. Chem. Soc. 101, 2815-2832.
Teo. B.-K., Antonio, M. R,, & Averill, B. A. (1983) J. Am. Chem. Soc. 105, 3751-3762.
Velthuys, B., & Kok, B. (1978) Biochim. Biophys. Acta 502, 211-221.
Vincent. J. B., & Christou, G. (1986) FEBS Lett. 207, 250-252.
Vincent, J. B., Christmas, C., Huffman, J. C., Christou, G., Chang, H.-R., & Hendrickson, D. N. (1987a) J. Chem. Soc., Chem. Commun., 236-238.
Vincent, J. B,, Chang. H.-R., Foiling, K., Huffman, J. Q Christou, G., & Hendrickson, D. N. (1987b) J. Am Chem Soc. 109, 5703-5711.
Witt, H. T,, Saygin, O., Brettel, K., & Schlodder, E. (1987) in Progress in Photosynthesis Research (Biggins, J., Ed ) Vol. 1, pp 523-531, Martinus Nijhoff, Dordrecht.
Yachandra, V. K., Guiles, R. D., McDermott, A., Britt, R D.. Dexheimer, S. L., Sauer, K,, & Klein, M. P. (1986a) Biochim. Biophys. Acta 850, 324-332.
Yachandra, V. K., Guiles, R. D., Sauer, K,, & Klein, M. P (1986b) Biochim. Biophys. Acta 850, 333-342.
Yachandra, V, K., Guiles, R. D., McDermott, A. E., Cole, J. L., Britt, R. D,, Dexheimer, S. L,, Sauer, K., & Klein, M P. (1987) Biochemistry 26, 5974-5981.
fiU
Mechanisms of Mutagenesis by the Vinyl Chloride Metabolite Chloroacetaldehyde. Effect of Gene-Targeted in Vitro Adduction of Ml3 DNA on DNA Template Activity in Vivo and in Vitro*
J. Steven Jacobsen and M. Zafri Humayun* Department of Microbiology and Molecular Genetics, University of Medicine and Dentistry of New Jersey--New Jersey Medical
School. 185 South Orange Avenue, Newark, New Jersey 07103
Received May 23, 1989; Revised Manuscript Received August 28, 1989
abstract: 2-Chloroacetaldehyde (CAA), a metabolite of the carcinogenic industrial chemical vinyl chloride, reacts with single-stranded DNA to form the cyclic etheno lesions predominantly at adenine and cytosine. In both ethenoadenine and ethenocytosine, normal Watson-Crick hydrogen-bonding atoms are compromised. We have recently shown that CAA adduction leads to efficient mutagenesis in Escherichia coli predominantly at cytosines, and less efficiently at adenines. About 80% of the mutations at cytosines were C-to-T transitions, and the remainder were C-to-A transversions, a result similar to that of many noninstructional DNA lesions opposite which adenine residues are preferentially incorporated. It is widely believed that noninstructional lesions stop replication and depend on SOS functions for efficient mutagenesis. We have examined the effects of in vitro CAA adduction of the lacZa gene of phage M13AB28 on in vivo mutagenesis in SOS(UV)-induced E. coli. CAA adduction was specifically directed to a part of the lacZ sequence within M13 replicative form DNA by a simple experimental strategy, and the DNA was transfected into appropriate unirradiated or UV-irradiated cells. Mutant progeny were defined by DNA sequencing. In parallel in vitro experiments, the effects of CAA adduction on DNA replication by E. coli DNA polymerase I large (Klenow) fragment were examined. Our data do not suggest a strong SOS dependence for mutagenesis at cytosine lesions. While adenine lesions remain much less mutagenic than cytosine lesions, mutation frequency at adenines is increased by SOS. SOS induction does not significantly alter the specificity of base changes at cytosines or adenines. The in vitro replication results show that these lesions create kinetic pause sites rather than absolute replication blocks. The simplest interpretation of the specificity of base changes is that the mechanisms of base incorporation opposite these lesions are analogous to those opposite the canonical noninstructional lesions. The high efficiency of mutagenesis opposite cytosine lesions without the aid of induced levels of SOS functions suggests that all DNA lesions lacking normal Watson-Crick base-pairing ability may nevertheless not block replication. In addition, the relatively nonmutagenic bypass of adenine lesions focuses attention on the need to understand mechanisms of error avoidance in the absence of normal base-pairing information.
ASX 00004964
Following the description of the formation of cyclic DNA
lesions by the mutagen glyoxal (Shapiro, 1969), a large number of chemical carcinogens have been shown to induce such lesions. For example, 2-chloroacetaldehyde (CAA),1 a metabolite of vinyl chloride (VC), predominantly reacts with unpaired adenine and cytosine residues to form l.M-etheno-
*This work was supported by USPHS Grant CA47234. M.Z.H. is the recipient of USPHS Research Career Development Award CA00907.
Author to whom correspondence should be addressed-
adenine (A) and 3,7V4-ethenocytosine (rC) which are the final stable lesions derived from the dehydration of the initially
1 Abbreviations: CAA, chloroacetaldehyde: VC, vinyl chloride; A, adenine; C, cytosine; T, thymine; G, guanine; ds, double stranded; ss, single stranded; polKk), Escherichia coli DNA polymerase I large (Klenow) fragment; iC, 3,,V-cthenocytosine; <C-H20, hydrated 3_V*ethenocytosine |3,,V*-(hydroxyethano)cytosine]; (A, !,A*-ethenoadenine; tA-HjO. hydrated I ,A*-ethenoadenine [l,,'V4-(hydroxyethano)adenineJ; UV. ultraviolet.
' Mechanisms of Mutagenesis by Chloroacetaldehyde
\
FIGURE I: Structures of eC and tA and their respective hydrated intermediate forms, (C-HjO and tA-H20. formed hydroxyethane intermediates A-H20 and (C-H20 (Figure 1). However, the biological impact of cyclic DNA lesions, especially their in vivo mutagenic potential, has been largely unexplored. Using a high-efficiency mutation detection system based on the lacZa sequence of coliphage Ml3, we have recently provided evidence that cytosine lesions inflicted by CAA (likely ethenocytosines) are highly mutagenic (Ja cobsen et al., 1989).
Mutagenic DNA lesions are usually classified into misin-
vrstructiona! and noninstructional categories on the basis of rported mechanisms of mutagenesis. Misinstructional le sions, exemplified by DNA uracil (a product of cytosine de amination) and (Amethylguanine, are assumed to cause mutations by simple miscoding and require only the normal DNA replication machinery. Noninstructional lesions, exem plified by abasic sites, UV photoproducts, and bulky chemical adducts, are believed to block DNA replication due to either a lack of template instruction or steric hindrance. Bypass of these lesions, at least in Escherichia coli, is presumed to require the intervention of induced genetic functions (the SOS path way). DNA sequence analysis of mutations induced by a few examples of noninstructional lesions suggests that, irrespective of the chemical diversity of these SOS-dependent lesions, adenines are most frequently incorporated opposite these le sions, followed by thymines and, much less frequently, other bases [sec, e.g., Kunkel (1984)]. Observation of a similar preference for adenine misinsertion opposite noninstructional lesions in defined in vitro systems has led to the hypothesis that this preference may be an inherent property of many DNA polymerases confronted with a site devoid of accessible template function [see, e.g., Rabkin and Strauss (1984)]. The specificity of misinsertion opposite CAA-induced cy tosine lesions (adenine, 80%; thymine, 20%), as well as other considerations, have led us to previously suggest that rC lesions might act as noninstructional lesions. Paradoxically, even though tA lesions were present at about the same level as tC lesions, mutations at adenine lesions were an order of mag nitude less frequent than at cytosines. Nevertheless, A-to-T 'fansversions predominated the few mutations at adenines, ggesting that these lesions were also acting as noninstruc tional lesions. Since noninstructional lesions should show a marked SOS dependence for mutagenesis and be able to act
Biochemistry, Vol. 29. So. 2. 1990 497
as replication stop sites, CAA-induced lesions should show these properties. In this paper, we describe in vivo and in vitro experiments to examine these issues.
Materials and Methods
Bacterial and Phage Strains, The suppressor-less . coli Kl2 strains KH2A (ucrA~) and KH2AM (KH2A carrying the mucAB* plasmid pGW270) were constructed as described by Sambamurti et al. (1988). The amber-less phage Ml 3AB28 was constructed as described (Sambamurti et al., 1988). AB28 contains the . coli /acZa-complementing fragment and yields dark blue plaques on suppressor-less . coli cells on appropriate indicator plates.
DNA Preparation and CAA Adduction. The preparation of AB28 ss DNA was as described (Rcfolo et al., 1987). Gapped duplex DNA was constructed as described in detail elsewhere (Jacobsen et al., 1989). In this DNA, a 178 nu cleotide long segment of the minus strand (within the lacZa gene) has been removed so as to leave the plus strand sequence between bases 6236 and 6413 in a single-stranded confor mation (see Figure 2). The preparation of CAA-treated DNA, adduct characterization, and quantitation have been described (Jacobsen et al., 1989). Briefly, the modification procedure involved a 15-min exposure of gapped duplex DNA at room temperature to 71.6 mM CAA, followed by DNA isolation and incubation at 80 C for 60 min to dehydrate the etheno lesions. This procedure resulted in an average of 2.1 and 3.4 tC and tA lesions, respectively, per DNA molecule. Mocktreated DNA (0 CAA) was prepared by the same protocol, including the post-"modification'' 80 oC/60-min incubation.
SOS Induction and Transfection Procedures, Procedures for transfection of CAA-treated and mock-treated (no CAA) DNAs into unirradiated . coli cells, plating of infectious centers, and determination of the total number of plaques, as well as the number of lacZ~ plaques, were as described (Ja cobsen et al,, 1989), Transfection efficiency of mock-treated gapped DNA averaged 10* plaque-forming units/ug and was close to that observed for untreated AB28 supercoiled ds DNA. For examining the effect of SOS induction, . coli KH2AM (or KH2A) cells were irradiated with ultraviolet (UV) light to induce the expression of SOS functions. Parameters for UV irradiation were as described (Sambamurti et al., 1988; Bennett et al., 1988).
Mutant Selection, DNA Isolation, and DNA Sequence Analysis. Mutant phage were identified as those exhibiting a light blue or colorless plaque phenotype in comparison to the dark blue phenotype of wild-type AB28. Plaque pheno types were confirmed essentially as described (Kunkel, 1984). Mutant phage ss DNA preparation and sequencing procedures were as described (Refolo et al., 1987).
In Vitro DNA Replication Analysis. Phage M13mp8 (Messing & Vieira, 1982) ss DNA or the AB28 gapped duplex DNA was subjected to CAA modification essentially as de scribed for gapped duplex DNA (Jacobsen et al., 1989). Adducted mp8 DNA was incubated for 1 h at either 0 or 65 C before use. Procedures for analyzing the template char acteristics of CAA-treated and mock-treated (no CAA) ss DNA were essentially as described previously (Jacobsen et al., 1987). Briefly, an appropriate oligonucleotide primer was annealed to the ss DNA template, and elongation was initiated at 37 C by . coli DNA polymerase large "Klenow" fragment [poll(k)] in the presence of either 8 mM MgCI2 or 0.5 mM MnCl2. Elongation products were radiolabeled by a 15-min pulse (2.14 mM [o-12P]dATP, 200 Ci/mmol, and 5 gM each of dGTP, dTTP, and dCTP) followed by a 15-min chase (50 fiM of each dNTP). Primer elongation was terminated by
ASI 00004965
1
498 Biochemistry, Vol, 29. No. 2, 1990
Table 1: Effect of in Vitro CAA Treatment of Phage Ml3 AB28 Gapped Duplex DNA on Survival and Mutagenesis in SOS-Induced
and Uninduced Cells*
host strain* (A) KH2AM
(B) KH2A
initial CAA exposure* (mM)
0 71 6
0 71.6
plaques/pg of DNA (XIO6) [% survival]
-SOS
+SOS
5.2 [100] 1.4 [27.0] 7.7 [100] 2.6 [33.3]
4.5 [100] 2.2 [49.2] 3.3 [100] l.l [31.6]
% mutagenesis
-SOS +SOS
0.41 4.92 0.20 1.66
t.07 12.6 0.55 5.43
Data shown are averages of two transfections. *. coli strain KH2AM is uirA', niucAB*-. KH2A is ucrA~, mucB" (see test). 'After CAA exposure. DNA was isolated and subjected to incubation for I h at 80 "C for conversion of hydrated lesions into the stable dehydrated lesions as described (Jacobsen ct al,, 1989). The number of adducts per gapped duplex molecule is estimated as 2.1 and 3.4, respectively, for tC and iA [see Jacobsen et al (1989)).
adding a commercially available (New England Biolabs) formamide-containing "stop" buffer, and the labeled elongation products were fractionated by electrophoresis on high-reso lution 8% polyacrylamide-urea gels. Analysis of the repli cation-blocking potential of CAA lesions was studied by using the same CAA-modified gapped duplex used for in vivo mu tagenesis. The single-stranded portion of the DNA was rep licated with DNA poll(k) by extending the 3'-hydroxy! at the Bgl1 site (right boundary of the target, see Figure 2). The elongation products were radiolabeled by a 5*min pulse, and replicate samples chased for I, 3, 30, or 90 min. Elongation was terminated by rapid freezing on dry ice, and the products were digested with A/srll. which cuts the DNA once ap proximately 75 nucleotides 3' to the Bgl] site. The resulting labeled DNA products were fractionated on high-resolution gels as above alongside standard dideoxy sequence ladders.
Results
In order to examine the effects of SOS induction on CAAinduced mutagenesis in E. coli, we have devised a convenient strategy using the phage M13 lacZ based experimental system for the detection and characterization of forward mutations (Jacobsen et al,, 1989). This "gapped" duplex strategy, by exploiting the known single-strand preference of CAA, permits the specific adduction of a part of the lacZa gene sequence. Furthermore, since the lesions are in the ss DNA, mutations can be unambiguously assigned as those occurring at G, A, T, or C. Plating on appropriate E. coli lawn cells and media containing a lac inducer and a chromogenic substrate for /J-galactosidase permits the identification of wild-type (blue plaques) and mutant (colorless or light blue plaques) phage. The construction of phage MI3AB28 as well as the basis for phenotypic selection have been described elsewhere (Sambamurti et al., 1988).
Effect of CAA Adduction and SOS Induction on Surcical and Mutagenesis. Table 1A shows that CAA treatment of DNA under conditions yielding an average of 2 eC and 3 <A lesions within the target sequence enhances mutagenesis by an order of magnitude over the background in unirradiated cells. Upon SOS induction, the fold enhancement in muta genesis remains at about an order of magnitude over the corresponding background. Since SOS induction increases both background and CAA-induced mutagenesis by about 2.6-fold, there is very little Weigle mutagenesis. The relatively high background mutation frequency is probably the result of DNA manipulations required for construction of the gapped duplex DNA molecules and post-"modification" heat treat ment. SOS induction does not significantly enhance survival
Jacobsen and Humayun
Table II. Base Changes in Apparent Wild-Type Plaques Obiained by Transfection of CAA-Treated DNA*
sequence change
sequence change
mut no.* base no. from to mut no.* base no. from to
1
6322
CA
6388
CA
2 6391 c A
3
6249
cA
6346
cT
6347
cT
6383
cT
4
6251
AC
5
6242
GT
6
6344
CT
7
6340
CT
6344
CT
8
6239
GA
6338
CA
9
6340
CT
10
6238
CT
11
6313
GC
6340
CT
12
6341
AG
6361
CA
13
6249
CT
6289 c T
14
6396
CT
15
6380
GA
16
6258
GA
17
6364
AT
18
6238
CT
19
6316
AT
6346
CT
20
6396
CT
21
6238
CT
22
6292
CT
23
6347
CT
24
6238
CT
6246
CA
25
6238
CA
6289
CT
26
6251
AT
6367
AT
6405
GC
27
6401
AT
*AII identified nucleotide changes were at sequences within the sinje-stranded portion of the gapped duplex molecule (Figure 2). Mutant No.
of CAA-treated DNA. It is difficult to assess the significance of the marginal survival effect observed (27% to 49% in Table IA), because replicate transfections often show a 2-5-fold variation in survival (e.g., Kunkel (1984), Refolo et al. (1987), and Sambamurti et al. (1988)].
The gene products of UmuD, UmuC, and RecA are believed to be required for SOS-dependent mutagenesis in E. coli. MucA and MucB (MucAB) are inducible plasmid encoded functional homologues to cellular UmuDC proteins and are frequently used in short-term assays to amplify the mutagenic response to certain DNA-damaging agents. For example, McCann et al. (1975) used an excision-deficient mucAB* (pKMIOl) strain of Salmonella to detect the mutagenic ac tivity of CAA. Table IB shows that fold enhancements in CAA mutagenesis in SOS-induced as well as uninduced KH2A cells lacking MucAB remain close to I order of magnitude. These results suggest that CAA mutagenesis does not require induced levels of SOS functions.
DNA Sequence Analysis of Background and CAA-fnduced Mutations. Figure 2 represents the DNA sequence of the single-stranded gap region and indicates the nucleotide changes (lower case letters) delected in 30 background mutants (shown above the sequence) and in 49 CAA-induced mutants (shown below the sequence) obtained under SOS-induced conditions. We have previously observed that the gene-targeted adduction strategy employed results in a large number of phenotypically silent (lacZ+) base substitutions (Jacobsen et al., 1989). To determine the true CAA-induced mutation frequency in SOS-induced cells, 40 plaques displaying a blue (lacZ+) phenotype were randomly picked from the same plates from which mutant plaques had been picked. Sequence analysis of the 40 apparent lacZ+ phage revealed that 27 (67.5%) contained one or more base substitution events (Table II).
The data in Figure 2 and Table II are summarized in Table III to show the classes of mutations as well as the specificity of the base changes. Table 111B shows that, out of 50 CAAinduced lacZ~ mutants sequenced, 49 (98%) suffer one or more substitutions. In addition, 3 of the 49 also contain a single
ASI 00004966
Mechanisms of Mutagenesis by Chloroacetaldehyde
Biochemistry, Vol. 29, No. 2, 1990 499
t i/S
6226
6216
. 6246
62S6
. 6266
6276
6266
( I 01 CC I ( I M I c tl TI CC [tG ( C 1 ( t G 1 [ GCC ! ( C 1 C C CMC [ 1 I t OCU I tottt I t t II T T C
TAC1CGTACTAATGCTTAA............................................
t/6 . -*B/1 " t/1 t/IS c/15
t/23 t/12 . t/4
t/13
.t
. i/1 . * -/21"
t /3
.1
. t/7 c/21
.t
. t/23
. t/13
t/12
. t/15
"/II"
t
t
t
t/*
- H /7
c t/2 . t
t t
t/2 . t/2
t t/4
t/3 c/8 t/6 -. H ./g " c
t/3 c.
6296
6306
. 6316 .
6326
6336 .
6346
6356 .
C G T C G T G A C T G G C A A A A C C C T GGCGI TACCCAAC T T A A T CGCCTTGCACCACAT C C C C C T T T c c c c
t/21 t/27 t
r~ 9/27
. . t/19 . . t/20 . t/25 , t/26 t/19
t/7 . t/5 t . i/17
t/5
t/lfl t/24
t/6 t . t/23 . t/14
i t 9 - t/3 . t/16 9 t/16 . t t/14 t/2 t/17 t/IS t/19
t/22 t/22 t/26
t/25
t . s/3 t t t/3 t/10
i/5
t . t/i
a/5 e
c t/i t/6
t/7
6366
. .6376
6386
6396
6406
6416
6426
C G T A A T A G C G A A G A G G C C C G C A C C G A T C C C c C T T C C C A A CACTTGCCCAGCCTGAATCCCCA A T G G
. , .. . , . .01/20. 9/10
t 9/2 t/n
c/4 t/18 t a/25
t/10
T TACCGCT TACC t/25
- .9/25 .t/9
t
C t/20
t/8
t/6 i/B
t/25
. t/24
9/17
t/9 t/27
a/12
figure 2: Pari of the phage AB28 lacZ DNA sequence indicating background and CAA-induced mutations. The plus strand is shown on top (written 5' to 3')- The single-stranded portion of the gapped duplex molecule was created by the procedures described elsewhere (Jacobsen et al.. 1989). All mutants sequenced were derived from the transfection of heat-treated (80 C) gapped DNA intoSOS(UV)-induced KH2AM
Ihplls. Background mutations are shown above the sequence. Multiple base changes occurring in the same mutant are identified by a common
number (c.g., base-change a/3 at position G6263 and change t/3 at C6346 both occur in the same mutant; similarly, the changes t/19 at A6316, t/19 at C6318, and t/19 at C6344 all occur in the same mutant). Dashes enclosed in double-quote marks (*-") indicate the four 1 bp deletions (positions C6292. G6337, and C6345 and a T in TTT6348-6350). CAA-induced mutants are shown below the sequence by use of conventions similar to those for background mutants. A plus enclosed in double-quote marks ("+*) indicates the single I bp insertion (positions GGGG6239-6242). Three CAA-induced mutations (C to A at 6195, part of mutant a/17; A to T at 6146, part of mutant t/28; and C to T at 6147, part of mutant t/28) mapping upstream of the initiating codon at 6217 are not shown but are included in the data in Table III. These three mutations are arbitrarily depicted as changing plus-strand bases.
Table 111: Summary of CAA-induced Base Changes in Apparent Wild-Type (lacZ*) Phage
DNA treatment (plaque phenotype)
(A) background (lacZ"/ (B) CAA induced (lacZ'/ (C) background (lacZ*/ (D) CAA induced (lacZ*/ (E) CAA induced (lacZ' and
lacZ*)1
ttl no. seqd"
34 50 20 40
no seq chng*
4 1 20 13
-126 bp mut'
1 0
mut with base
base substitution events'
events
A to
C to
G to
subst* XI X2 X3 X4 X6 ttl c g t a 8 a c t
26 19 5 2 0 0 35 0 0 4 2 0 6 4 6 13
49 22 16 7 3 1 93 0 3 13 7 5 55 2 4 4
0 27 16 9 1 1 0 41 1 1 5 8 0 20 3 2 1 0 76 38 25 8 4 1 134 1 4 18 15 5 75 5 6 5
Total number sequenced. `Number showing no sequence changes between bases 6130 and 6430. '126 bp deletions, representing recombinational background, are analogous to the "93 bp deletions* in M13mp2 (Kunkel, 1984: LeClerc et al,, 1984). The polylinker sequence in MI3AB28 adds 33 bp to the size of the deletion. ^Number suffering base substitutions. 'XI, X2, X3. X4, xfi, and ttl represent cither one, two, three, four, six, or total substitution events, respectively; sequence changes in the plus strand are shown. 1Data for phenotypical!) delected (lacZ') mutations (background and CAA induced; 80 'C treatment). 'Analysis of 20 randomly picked blue (lacZ*) plaques resulting from transfection of mock-treated, 80 C incubated gapped duplex DNA (see text). `Analysis of 40 randomly picked blue (lacZ*) plaques resulting from transfection of CAA-treated, 80 C incubated gapped duplex DNA (see text). 'Pooled data (sum of rows B and D) on base changes observed as a result of CAA treatment of DNA.
frame-shift event. These include two I bp deletions (at A6243 and A6285; see Figure 2) and a G insertion in the GGGG run at 6239-6242. The 34 background lacZ' mutants (Table II11 A) are comprised of fewer base substitutions (76%). One
Iutant has a single substitution at G6352 and a I bp deletion
t a TTT run at 6348-6350 (mutations marked with /8 in Figure 2). Three frame-shift mutants arise by the loss of a single base pair (C6292; G6337; and a C in the CCCCC run
at 6343-6347; Figure 2). The data in Table HID show that CAA-induced silent mutations have a specificity similar to that of phage with a lacZ' phenotype. Sequence analysis of 20 randomly picked lacZ+ plaques from control DNA plates showed that all 20 had the wild-type sequence (Table 111C). These data show that, under SOS conditions, CAA mostly induces base substitutions, a result similar to that obtained in unirradiated cells (Jacobsen et al,, 1989).
AS I
500 Biochemistry, Yol. 29. No, 2. 1990
Table IV: Analysis of CAA Enhancement of Mutagenesis with and without SOS Induction at A, C, G, and T Residues*
rel mutation frequency (1(T3)
background
CAA
all mutations at A at C at G at T
-SOS
4 0.37 0.73 2.9
+SOS
10.7 1.22 2.45 7.03
-SOS
575 34
520 21
+SOS
801 137 568 96
No mutations were detected at thymines; the "-SOS" data are from Jacobsen et al. (1989). The "+SOS" background mutation fre quency is from Table I. CAA-induced "unselected' mutation fre quency was derived from the data in Table HID (lacZ+, CAA in duced) as follows;
no. of plaques with base changes (27) -------------------------------------------------------- = 0.675
no. of plaques sequenced (40)
It should be noted, however, that the so-called "unselected" mutants (blue plaques) are in fact "reverse-selected" in the sense that light blue and colorless plaques (together constituting 12.6% of all plaques; Table I) arc omitted from this collection. Therefore, the true total mutation frequency (MF) is derived as follows:
MF = unselected MF (0.675) + selected MF (0.126) = 0.801
Relative mutation frequency (RMF) is the fractional mutation fre quency contributed by mutations occurring at G, A, T, or C to the MF, as shown for CAA-induced adenine targeted mutations:
no. of A-targeted substitutions (23)
RMF =
(MF (0.801)]
total substitutions (134)
(Note that substitution frequency should be used in the place of MF in this calculation; however, since an overwhelming majority of CAA-in duced mutations are substitutions, substitution frequency is taken to equal MF.) There are no significant specificity differences in the se lected and silent mutations: e.g., the fraction of C-targeted mutations in the selected pool (Table III8) is 67/93 = 0.72, and for the unse* lected mutations (Table HID), it is 28/41 = 0.68. Therefore, using the pooled sequence data simply serves to increase the number of total mutations without a significant effect on the fraction of the mutations occurring at each of the four bases. RMF values can also be calculated specifically for selected mutations (MF = 0.126; Table I), or specifi cally for silent mutations (MF = 0.675). Since the true mutation fre quency is determined by the number of mutants in the unselected as well as the selected pool as shown above, RMF figures for the pooled data are likely to be the most representative.
Specificity of CAA-induced Base Substitution Mutations. Table HIE shows that the majority of CAA-induced base substitution mutations are targeted to cytosines (71%), with fewer occurring at adenine (17%) and guanine (12%) residues, and with none occurring at thymines. The corresponding figures for background substitutions are C (23%), A (11%), and G (66%). Table 1IIE also shows that, of the 95 CAAinduced base changes affecting cytosines, 75 are C-to-T (79%) mutations, followed by 15 C-to-A (16%) and 5 C-to-G (5%) mutations. Of the 23 adenine-targeted substitutions, 13 (78%) are A-to-T, 3 are A-to-G, and 1 is an A-to-C substitution. These results, in comparison to those in uninduced cells (Ja cobsen et al., 1989), show that SOS induction does not sig nificantly alter the specificity of base changes induced by CAA.
Efficiency of CAA-induced Mutagenesis. Relative mutation frequency calculations (Table IV) show that CAA treatment increases overall mutagenesis as well as cytosine-specific mutagenesis to a similar extent in SOS-induced and uninduced cells. Substitutions at A and G residues are enhanced nearly 4-fold with SOS induction. However, in comparison to mu tation fold enhancement at cytosine (200-600-fold) and at adenine (100-fold), the increase at guanine (6-10-fold) is modest.
ASI 00004968
Jacobsen and Humay
Table V: Efficiency of Mutagenesis of CAA-induced Cytosine Lesions with and without SOS Induction
(A) surviving fraction (B) mutant fraction' (C) wild-type fraction (1 - B)r (D) corrected wild-type fraction (C X A)* (E) corrected mutant fraction (B X A)* (F) no. of mutants (selected and unselected) with
base substitutions (G) no. of mutants with single C-targeted
substitutions' (H) mutant fraction with single C-targeted events
[E X (G/F)] (I) mutagenic efficiency at single C-targeted events
(H/0.271/
-SOS*
0.300 0.575 0425 0.128 0.173 69
+SOS
0.49( 0.80: 0.19< 0.09: 0.39. 76
48 24
0.120 0.12 0.443 0.45
"-SOS' values are taken from indicated sources in Jacobsen el (1989). *"+SOS" values are taken from sources indicated in tl present paper. ' Fraction of survivors with (B) or without (C) ba changes. ^Wild-type (D) and mutant (E) fractions corrected for su vival. 'These figures were obtained by counting the number of m tants suffering only one C-targeted mutation [Figure 4 of Jacobsen al. (1989); Figure 2 and Table II]. fAt an average of 2 tC lesions p molecule, the Poisson fraction of DNA molecules suffering 0, 1,2, a> more than 2 fC lesions are 0.135, 0.271, 0.271, and 0.323, respective
As considered further under Discussion, mutagenic cytosi: lesions induced by CAA are probably <C (or <C-H:0). Fro the average number of tC lesions per DNA molecule, and t frequency of single-base substitution events at template residues, it can be argued that the efficiency of mutager bypass of single tC lesions is over 40% (Table V) in be SOS-induced and uninduced cells. These calculations shot be viewed as first approximations until the availability of d with site-specific eC lesions at a number of sites. Simi calculations for events at A residues (not shown) suggest tl mutagenesis at single (A lesions is inefficient (3%) in t absence of SOS induction but increases (to about 13%) SOS-induced cells. An interesting feature of the analysis Table V is that the corrected wild-type fractions (Table V 0.128 for -SOS and 0.098 for +SOS) are close to the Pois: fraction of molecules bearing zero tC lesions (0.135). Si: at 3 eA lesions (on the average per DNA molecule) the Pois: fraction of DNA molecules with zero <A lesions is less tl 0.05, these calculations imply that (A lesions make only a sn contribution to mutagenesis and contribute little to lethal
Effect of DNA Sequence Context on the Frequency c Specifcity of CAA-Specific Substitution Mutations Template C Residues. In order to analyze the effect of cleotide sequence on CAA-induced mutagenesis at cytosi: we have plotted the normalized relative mutation freque for each of the 14 NCN triplet sequences found (out of possible NCN sequences; Figure 3) in the ss DNA "gap", ` salient points derived from this analysis can be summari as follows. (I) Mutagenesis is detected at all triplet nucleo sequence environments represented within the ss DNA tai (although not necessarily at every occurrence of the s; triplet sequence), supporting the notion that CAA probr reacts randomly with all non-base-paired C residues [see cobsen et al. (1987)]. (2) Normalized data for C-to-N evi (Figure 3) or C-to-T and C-to-A events (data not shown) not significantly affected (Student's t test) by SOS induct (3) In general, 5'-C-Pu-3'sequences are associated with hi; frequencies of mutagenesis as compared to 5'-C-Py sequer
Effect of CAA Modification of Template DNA on Pri Elongation by poll(k). In order to determine whether C adducts stop or attenuate DNA synthesis, M13mp8 ss D was subjected to CAA modification, and DNA replicatio the modified template was initiated by primer extension ir
Mechanisms of Mutagenesis by Chloroacetaldehyde
Normalized relative mutation frequency (x to'3 )
10 20 30 40 50
Biochemistry, Vol. 29, No, 2, 1990 501
DIDEOXY CAA HOCK GATCxz^l
-US . |I8 3
e 100 a&e
Pu-C-Pu Pu-C-Py Py-C-Pu Py-C-Py
figure 3: Effect of SOS induction and nucleotide sequence envi
ronment on CAA-induced base substitution mutations at C lesions.
Normalized relative mutation frequencies (RMF) at each of the 14
NCN triplets found within the "gap" region (out of a possible 16)
are depicted. The normalized RMF was calculated from the RMF
for all mutations occurring in a particular triplet and the number of
mrrences of the triplet within the ss DNA gap sequence, as ilrated for the triplet 5'-ACA (+SOS conditions) in eq I and 2.
RMF(5'-ACA) =
C mutations in 5'-ACA (14)
ail events at C (93)
|RMF(C) (0.S68)) (=0.0855) (1)
RMF(5'-ACA) (0.0855) normalized RMF(5'-ACA) no. of 5'-ACA triplets (3) ^
(2)
The RMF(C) value is from Table IV. Grey bars represent -SOS data, and dark bars, +SOS data. The sequences 5'-TCT or 5'-TCA (asterisks) are not represented in the 178-base ss DNA gap. and therefore, no data are available for these sequence contexts. Panel A, effect of each individual base 5' or 3' to C; panel B, effect of purine (Pu) or pyrimidine (Py) flanking the C. In the above analysis, 28 unselected and 67 selected C-targetcd events distributed over 5) sites are analyzed. The rationale for pooling the sequence data is given in the footnote for Table IV. Because of the inclusion of the unselected mutations, the pool is taken to represent most of the base changes induced by CAA within the target sequence. It should be noted that, with or without the "normalization" (eq 2 above), the major conclusions regarding mutational hot spots as well as the lack of SOS effect remain the same.
presence of E. coli DNA poll(k), divalent cation, and nu cleotide precursors (Rabkin & Strauss, 1984). Labeled elongation products were analyzed on DNA sequencing gels. Preliminary results using this approach had shown that CAA-induced DNA damage creates pause sites at A and C residues (Jacobsen et al., 1987). It was observed that repli cation terminated predominantly one nucleotide before (3' to)
lesion, and less frequently opposite the lesion. Figure 4
that the substitution of MnJ+ for Mg2+ (known to be lagenic; Sirover & Loeb, 1976) causes qualitative changes in replication stops at A and C lesions. At certain sites, there is a shift in the replication pause site from one nucleotide before
figure 4: In vitro replication analysis of CAA-modified M13mp8 ss DNA template in presence of either MgJ+ or Mn2+. The lanes G. A, T. and C represent standard Sanger (dideoxyribonucleotide) se quence ladders. The relevant part of M13mp8 lacZ sequence (primer strand indicated by arrow) is shown. CAA-modified or mock-modified (no CAA control) ss DNA template was replicated with DNA poll(k) by primer extension in the presence of either 8 mM MgClj (Mg) or 0.5 mM MnClj (Mn) as described under Materials and Methods. The series of bands in CAA-modified ss DNA lanes represent po lymerase pause sites. In mock-treated DNA lanes, there are no visible bands, indicating lack of chain termination. (With mock-treated templates, high molecular weight DNA products traveling at the top of the gel are observed; data not shown.) The pause sites in CAAtreated DNA correspond to chain termination events opposite or, more frequently, one nucleotide before (3') template C or A residues. CAA-treated DNA replicated in the presence of MnClj shows fewer pause sites in comparison to the DNA replicated with MgClj. In the presence of MnC!2, several bands arc apparently completely bypassed (e.g., band one nucleotide 3' to template C-6264 at the bottom of the CAA/Mg lane), while at others, there is a "shift-up" effect, suggesting that terminations occurred more frequently opposite a template C or A residue (e g., compare terminations opposite template C-6236 in the CAA/Mg and CAA/Mn channels).
the lesion to a position opposite the lesion. At other sites, a complete bypass of the lesion is apparent. These effects, which appear to be influenced by the sequence context, are observed at several A and C residues. These results suggest that CAA adduction specifically creates pause sites at adenine and cy tosine residues and that poll(k) can incorporate bases opposite CAA-induced adenine and cytosine lesions in vitro.
To determine whether CAA-induced lesions introduced absolute replication blocks to polymerase or kinetic (tempo rary) pause sites, the ss DNA portion of CAA-modified gapped
duplex DNA (used in mutagenesis studies described above) was replicated by elongating the 3'-OH end of the comple mentary strand as described (see Materials and Methods and legend to Figure 5). Figure 5 shows that, with Mg1* as a cofactor, polymerase pause sites are observed initially (1-3
ASI 00004969
502 Biochemistry, Vol. 29, So. 2, 1990
DIDEOXY
Hafa
6 A T C 1 3 30 90 1 3 30 90
figure 5: Effect of time of incubation on the in vitro poll(k) pause sites on CAA-treated ss DNA template contained within the gapped duplex. For these experiments, the ss DNA portion of CAA-modifted gapped duplex DNA was replicated from the 3'-hydroxyl terminus present opposite template G64I4 in the presence of poll(k), nucleotide precursors, and either Mg2+ (Mg) or MnI+ (Mn) as described; see Materials and Methods. The numbers I, 3, 30, and 90 indicate the number of minutes the elongation reaction was allowed (following an initial 5-min pulse) before termination, G, A, T, and C are standard Sanger sequence ladders. A relevant part of the ss DNA sequence is shown (primer strand identified by an arrow). The gradual decrease in relative intensity of bands corresponding to pause sites with time is observable at several sites in the Mg lanes and is much more pronounced in the Mn lanes for any given time interval. For an explanation of the arrow, right-pointing triangle, and bracket, see text.
min) at almost all A and C residues. Some of the pause sites disappear at longer time intervals (30-90 min), indicating that these lesions temporarily stall DNA synthesis and are not absolute blocks. The figure also suggests that MnJ+, as com pared to MgH, increases the efficiency of (i.e., decreases the time required for) in vitro translesion synthesis.
The data in Figure 5 suggest that bypass synthesis may occur more readily at some sequences than at others. The example marked by the arrow corresponds to a template C residue in the sequence 5'-ACA-3' (6339-6341; see Figure 3), known to be a mutational hot spot in vivo. (This effect was also observed at a second 5'-ACA-3' site at 6285-6287; data not shown.) Despite this apparent correlation between in vitro bypass and in vivo mutagenesis, in vitro translesion bypass may not accurately predict in vivo mutagenesis. Figure 5 shows that equivalent pause sites are initially formed at all five bases in the CCCCC run at 6343-6347 (marked by the bracket). However, at 90 min (Mg2+) the blocks at 6343,6345, and 6347 (corresponding to the first, third, and fifth bracketed bands, respectively, from the top) are diminished in comparison to those at 6344 and 6346 (corresponding to the second and fourth bracketed bands, respectively). Nevertheless, in vivo mutations were observed at position 6344 (closed triangle. Figure 5; also see Figure 2), but not at the lesions that appear to be more readily bypassed in vitro. This observation raises the possibility that guanine may be incorporated opposite C lesions at certain sites in vivo, resulting in error avoidance. In the above analysis, we have assumed that equivalent pause sites imply equivalent (random) adduction, an assumption that remains to be tested. Finally, caution is required in extrap
Jacobsen and Humayun
olating in vitro replication results obtained with a model po lymerase such as poll(k) to in vivo mutagenesis, which is undoubtedly much more complex.
Discussion
Mutagenic DNA Lesions Induced by CAA. The prepon derance of evidence, reviewed briefly by Jacobsen et al. (1989), indicates that the mutagenic lesions induced by CAA are etheno derivatives, notably ethenocytosine. It is conceivable that a small fraction of the C-to-T mutations, but not the approximately 20% mutations changing C to A or G, are caused by simple deamination of cytosine during the in vitro modification procedures. However, the following considera tions suggest that the cytosine deamination to yield DNA uracil is unlikely to play a significant role in the observed mutagenesis. (1) Hall et al. (1981), who specifically looked for cytosine deamination during CAA treatment, did not find significant deamination. (2) The postlabeling procedure used by us can detect the formation of deoxyuridine at significant levels, but none was detectable by this method. (3) In order to increase the sensitivity of detection, we have examined deamination of deoxycytidine 5'-monophosphate subjected to the various modification and incubation conditions used here (data not shown). CAA treatment by itself resulted in little deamination above the background. The postmodification 80 C/l-h incubation, as expected, did result in detectable deamination, which, however, constituted a small fraction of eC formed during the procedures. No detectable deamination occurred after a 0 C/l-h incubation. (4) Jacobsen et aL (1989) have shown that the frequency and specificity of mu tations obtained by transfection of DNA subjected to 0 C/l-h incubation do not significantly differ from those of 80 C/l-h incubated DNA. (5) The background mutations analyzed here (Table JIIA) and elsewhere (Jacobsen et al., 1989) were ob tained by transfecting mock-treated DNA subjected to the 80 C/l-h incubation. Nevertheless, C-to-T mutations are a smaller fraction of background mutations as compared to mutations at guanines. Even if it is unlikely that simple cy tosine deamination could account for a significant fraction of mutagenesis, random-adduction experiments of the type de scribed here always leave open the possibility that other, as yet unidentified, primary or secondary cytosine lesion(s) may be responsible for the observed mutagenesis. Nevertheless, this work does establish that in vitro CAA treatment of ss DNA results in a large increase in SOS-independent muta genesis specifically at cytosines and that this in vivo muta genesis correlates with a CAA dose-dependent formation of C lesions in DNA.
Role of Cytosine Lesions in Mutagenesis Induced by Vinyl Chloride. Vinyl chloride, a major industrial chemical produced in large quantities around the world, is a well-known carci nogen. Nevertheless, the DNA lesions responsible for the genotoxicity of vinyl chloride are not known. Vinyl chloride is known to be metabolized in vivo to 2-chloroethyIene oxide, which can spontaneously and rapidly rearrange to CAA (Guengerich et al., 1979). The major DNA adduct induced by vinyl chloride (via chloroethylene oxide) in vitro and in vivo is 7V-7-(2-oxocthyl)guanine (oxet-G; Laib et al., 1981). In vitro reaction of metabolically activated vinyl chloride with DNA (Laib et al., 1981) and RNA (Laib et al., 1977, 1978) also yields eC and cA; reaction with DNA also yields N2,!ethenoguanine (Oesch & Doerjer, 1982). After the original demonstration of eC and eA in the livers of rats subjected to low-level vinyl chloride exposure over a period of time (Green & Hathway, 1978), this finding could not be confirmed until recently. Since induction of eC and (A lesions in DNA
ASI 00004970 "
Mechanisms Qf Mutagenesis by Chloroacetaldehyde
probably occurs only in single-stranded regions, these lesions arc expected to be present at very low levels, making their detection and quantitation in genomic DNA difficult. How ever, very recently. Eberle et al. (1989), using a specific monoclonal antibody, have detected and quantitated eC and <A
Iki liver and lung DNA of rats exposed to vinyl chloride. w In bacterial reversion systems, 2-chlorocthylene oxide in
duces mostly G-C to A*T transitions, and fewer A-T to T-A transversions (Barbin et al.. 1985a), and this mutagenesis is largely SOS-independent (Barbin et al., 1985c). Therefore, both the mutational specificity and the lack of SOS dependence of 2-chloroethylene oxide are strikingly similar to those of CAA (this work and Jacobsen et al. (1989)] and can be readily explained by postulating that the significant mutagenic lesions formed by chloroethylene oxide exposure were in fact the same cytosine lesions as those formed by CAA in vitro. However, since the reversions could have arisen by DNA lesions at either a guanine or a cytosine, it is important to evaluate whether the oxet-G adduct could lead to this specificity. Because of a lack of in vitro miscoding properties, oxet-G is believed to be nonmutagenic (Barbin et al., 1985b). Since guanine N7 adduction by bulky chemicals can lead to mutagenesis [e.g., Sambamurti el al. (1988)], it is premature to dismiss a mu tagenic role for oxet-G. Any hypothesis implicating the guanine adduct (e.g., an abasic site intermediate), however, must account for both the mutational specificity as well as the SOS independence in bacterial test systems.
Possible Mechanisms of Mutagenesis by tC. The specificity of mutagenesis at cytosines suggests that C lesions are treated as noninstructional lesions by the polymerase. Such a hy pothesis is consistent with the expected loss of normal hy drogen-bonding capability in tC (Figure I) and with the ob-
vation that DNA containing hydrated as well as dehydrated lesions yields similar patterns of mutagenesis (Jacobsen et al., 1989). This hypothesis must, however, account for the SOS-independent, highly efficient mutagenic bypass of cyto sine lesions and the observation that these lesions cause tem porary pause sites in vitro rather than absolute blocks expected of "true* replication-blocking lesions such as aflatoxin B, (Refolo et al., 1985; Jacobsen et al., 1987). The apparent inconsistency can be explained by assuming that eC lesions do not impede the progression of the polymerase because their structural properties permit the maintenance of the integrity of the helix in the absence of proper hydrogen bonding. Two caveats need to be considered here; First, the E. coli DNA polymerase utilized in mutagenic DNA replication past the DNA lesions in the gapped DNA in vivo is not known and can be either pol111, poll, or even polll. Therefore, the observed SOS independence may have other explanations. Second, even though inferences on in vivo mutagenic mechanisms have been made in the literature on the basis of simple in vitro replication systems using model polymerases such as poll(k), the validity of such extapolation remains to be investigated.
Finally, it is possible that eC has multiple miscoding prop erties which account for the various types of cytosine-targeted mutations observed. However, until convincing miscoding schemes become available, it may be useful to consider these lesions to be a special class of highly mutagenic noninstruc tional lesions.
Mutagenesis and Error Avoidance at tA Lesions. Jacobsen al. (1989) have previously suggested that eA lesions exist an equilibrium between an "unfavorable* and a "favorable" configuration. In the unfavorable state, they act as mutagenic noninstructional lesions responsive to SOS functions. This state can be imagined as a helix-distorting conformation. While
Biochemistry. Vol. 29, No. 2, 1990 503
the present data do show' that mutagenesis at adenines (unlike that at cytosines) is SOS-responsive, the facile, error-free, bypass of a majority of these lesions is difficult to explain.
It is possible that a significant fraction of tA lesions are spontaneously or enzymatically reversed before replication in vivo. This might also account for the observation that in vitro replication pause sites are found at both adenines and cytosines (Figures 4 and 5), and yet adenine lesions do not appear to cause significant lethality. However, such a hypothesis must account for the following observations: (1) the extensive literature on these derivatives, which were extensively used as enzyme substrate analogues for adenine (Leonard, 1984), suggests that these are stable lesions; (2) repair enzymes are generally believed not to act on ss DNA and therefore may not be able to eliminate A lesions before replication; (3) in the present system (as for other ss DNA phages), a strand discontinuity caused by excision repair is expected to be lethal, whereas tA lesions appear to contribute little to lethality. In order to account for the error-free bypass of these lesions, we have previously suggested that tA lesions in the alternative "favorable" conformation are capable of templating the in corporation of thymine at a high efficiency (Jacobsen et al., 1989). Whether such "templating" involves stabilization of an tA-T "base pair" by novel hydrogen bonds, or repesents a novel helix-stabilizing mechanism involving stacking and other hydrophobic forces, remains to be investigated.
Effect of Sequence Environment on CAA-Induced Muta genesis. The availability of the DNA sequences of a large number of phcnotypically selected as well as silent mutations presented in this paper and elsewhere (Jacobsen et al., 1989) offers an opportunity to examine the effects of sequence context on the mutational process. There is at present no evidence that the sequence environment influences the re activity of CAA with unpaired adenine or cytosine. (This is distinct from the decreased reactivity of adenine or cytosine in base-paired regions such as the stems of hairpins; see Ja cobsen el al. (1987)]. Even though interpretation of such data is not unambiguous. Figure 5 shows that initial pause sites for a model polymerase at many cytosine sites are equivalent, implying random adduction. The distribution of mutations is distinctly nonrandom (Figures 2 and 3), a phenomenon echoed as variability in the bypass efficiency of these lesions during in vitro replication by a model polymerase (Figure 5). The in vivo observations presented here imply that the mu tagenic potential of a DNA lesion is strongly modulated by the sequence environment. In that sense, an identical chemical lesion can in fact give rise to a number of distinctly different premutagenic lesions.
Acknowledgments
We thank X. Luo and J. Callahan for assistance with DNA sequencing and K. Sambamurti for strains and helpful dis cussions.
Registry No. CAA, 107*20-0.
References
Barbin. A.. Besson, F., Perrard, M.-H., Be're'ziat, J.-C., Kaldor, J., Michel, G., & Bartsch, H. (1985a) Mutat. Res. 152, 147-156.
Barbin, A., Laib, R. J., & Bartsch, H. (1985b) Cancer Res. 45, 2440-2444.
Barbin, A., Tenenbaum, L,, Toman, Z., Radman. M., & Bartsch, H. (1985c) Mutat. Res. 152, 157-159.
Bennett, C. B., Luo. X., Refolo. L. M., St Humayun, M. Z. (1988) Mutat. Res. 202, 223-234.
ASX 00004971
504 Biochemistry 1990, 29, 504-511
Eberle, G.. Barbin, A,, L&ib, R. J., Ciroussel, F,, Thomalc, J., Bartsch, H., & Rajewsky, M. F. (1989) Carcinogenesis 10, 209-212.
Green, T., & Hathway, D. E. (1978) Chem.-Biol. Interact, 22, 211-224,
Guengerich, F. P., Crawford, W. M,, Jr,, & Watanabe, P. G. (1979) Biochemistry 18, 5177-5182.
Hall, J. A,, Saffill, R., Green, T., & Hathway, D. E. (1981) Carcinogenesis 2, 141-146.
Jacobsen, J, S., Refolo, L. M., Conley, M. P,, Sambamurti, K. , & Humayun, M. Z. (1987) Mutat. Res. 179, 89-101.
Jacobsen, J. S., Perkins, C. P., Callahan, J. T., Sambamurti, K., & Humayun, M. Z. (1989) Genetics 121, 213-222.
Kunkel, T. A. (1984) Proc. Natl. Acad. Sci. VS.A. 81, 1494-1498.
Laib, R. J,, & Bolt, H. M. (1977) Toxicology 8, 185-195. Laib, R. J., & Bolt, H, M. (1978) Arch. Toxicol. 39, 235-240. Laib, R. J,, Gwinner, L. M., & Bolt, H. M. (1981) Chem.-
Biol. Interact. 37, 219-231.
LeClerc, J. E., Istock, N. L,, Saran, B. R,, & Allen, R. (1984) J. Mol. Biol. 180, 217-237.
Leonard, N. J. (1984) CRC Crit. Rev. Biochem. 15,125-199, McCann, J,, Simmon, V,, Streitweiser, D., & Ames. B. (1975)
Proc. Natl. Acad. Sci. US.A. 72, 3190-3193. Messing, J., & Vieira, J. (1982) Gene 19, 269-278. Oesch, F., & Doerjer, G. (1982) Carcinogenesis 3, 663-665. Rabkin, S. D.. & Strauss, B. S. (1984) J. Mol. Biol. 178
569-594. Refolo, L. M., Conley, M. P., Sambamurti, K., Jacobsen, J.
S., & Humayun, M. Z. (1985) Proc. Acad. Natl. Sci. V S.A. 82, 3096-3100. Refolo, L. M., Bennett, C. B,, & Humayun, M. Z. (1987) J. Mol. Biol. 193, 609-636. Sambamurti, K., Callahan, J., Luo, X., Perkins, C. P., Ja cobsen, J. S., & Humayun, M. Z. (1988) Genetics 120 863-873. Shapiro. R. (1969) Ann. N.Y. Acad. Sci. 163. 624-630. Sirover, M. A., & Loeb, L. A. (1976) Science 194, 1434-1436,
Proton NMR Studies of Transforming and Nontransforming H-ras p21 Mutants
lime Schlichting, Jacob John, Mathias Freeh, Pierre Chardin,* Alfred Wittinghofer, Herbert Zimmermann,* and Paul Rosch*
Department of Biophysics, Max-Planck-Institut for Medical Research, Jahnstrasse 29, D-6900 Heidelberg. West Germany Received April 18, 1989; Revised Manuscript Received August 24, 1989
ABSTRACT: One- and two-dimensional nuclear magnetic resonance spectroscopy (ID and 2D NMR) and site-directed mutagenesis were used to study the influence of mutations on the conformation of the H-ros oncogene product p21. No severe structural differences between the different mutants, whether they were transforming or nontransforming, could be detected. Initially, selective incorporation of 3,5-deuterated tyrosyl residues into p2l and 2D NMR were used to identify the resonances representing the spin systems of the imidazole rings of the three histidyl residues in the protein, of six of the nine tyrosyl rings, and of four of the five phenylalanyl rings. The spin systems of the phenyl rings of Phe28. Phe78, and Phe8: could be assigned by using mutant proteins, since no severe structure-induced spectral changes in the aromatic part of the spectra of the mutant proteins were detected. Sequence-specific assignments of the histidine imidazole resonances could be obtained by comparison of the distance information obtained by nuclear Overhauser enhancement spectroscopy (NOESY) experiments with the crystal structure. The change in the chemical shift values of the HI' proton and the a-phosphate of the bound GDP in the NMR spectra of the p21(F28L) mutant and the 28-fold increase in the GDP dissociation rate constants of this mutant suggest a strong interaction between Phe18 and the p21-bound nucleotide. In solution, the p21-bound GDP-Mg2+ has an anti conformation, and the phenyl ring of Phe28 is close to the ribose of the bound GDP*Mg2+.
JLhe products of the ras gene family are highly related pro teins of molecular weight 21000 termed p21. They have chain lengths of 189 amino acids; their sequences are identical for the N-terminal 80 amino acids and over 85% identical up to amino acid 164/165. The C-terminal 25 amino acids are very divergent, except for the Cys-A-A-X-OH motif at the end of the chain (A being an aliphatic residue). Normal (cellular) ras genes acquire transforming properties by single point mutations within their coding sequences, ras genes carrying these mutations have been detected in a significant fraction of human cancers as well as in experimentally induced animal tumors [for a recent review, see, e.g., Barbacid (1987)]. p21
'Present address: INSERM U-248 Facultf de Medecine Lariboisiere--Saint-Louis, 10, avenue de Verdun, 75010 Paris, France.
* Present address: Molecular Crystal Group, Max-Planck-lnstitut for Medical Research, Molecular Crystal Group, Jahnstrasse 29, D-6900 Heidelberg. West Germany.
proteins are thus believed to play an essential role in cellular growth and/or development (Bishop. 1983). They bind gua nine nucleotides with high affinity and specificity and exhibit low GTPase activity (Feuerstein et al., 1987; Gibbs et a!., 1984; McGrath et al,, 1984; Sweet et al.. 1984; Manne et al., 1985). Due to their significant sequence homology and biochemical similarities to guanine nucleotide (G) binding proteins such as transducin and the bacterial elongation factor Tu (EF-Tu),1
1 Abbreviations: ID, one dimensional; 2D. two dimensional; COSY, correlated spectroscopy; DQF, double quantum filtered; DSS. sodium 2,2-dimethyi-2-silapentanesulfonate; EDTA, ethylenediaminetetraacetic
acid; EF-Tu, bacterial elongation factor EF-Tu; GMP. guanosine 5'monophosphate; GDP, guanosine 5'-diphosphate; GTP, guanosine 5'triphosphate; H-ras, Harvey ras; NOE, nuclear Overhauser effect; NOESY, nuclear Overhauser enhancement spectroscopy; NMR, nuclear magnetic resonance; p2le, cellular H-ras p2l: p2l,, viral H-ras p2l (GI2R, A59T); pK,, apparent pAT value; TPPI, time-proportional incre
mentation method.
ASI 00004972
!